This article was first published in the Financial Post on January 13, 2008.
In the old days, stock exchanges had a monopoly on trading listed stocks. Not any more. These days, electronic platforms known as “alternative trading systems” (ATSs) provide investors with a variety of trading forums. Some of these, like BlockBook, Liquidnet and MATCH Now, exist solely to cross large (mostly institutional) blocks of stock. But others, like Pure Trading, Alpha, Chi-X Canada and Omega, serve the entire investment community — trading some or all of the TSX and TSX-V listed stocks.
By introducing competition to stock trading, ATSs have already lowered the cost and increased the speed and efficiency with which stocks are traded. Nonetheless, the ATS phenomenon is still in its infancy in Canada — and so is the regulatory framework. While the regulatory apparatus mostly works very well, some recent trading practices have arisen that materially compromise the fundamental principles of price discovery and liquidity that lie at the very core of a modern trading system.
Price discovery involves the determination of a suitable trading price for a given security. Liquidity occurs when investors can easily (and at reasonable cost) find a counterparty with whom to trade.
Price discovery works best when all traders have ready access to all other orders submitted for execution. Likewise, the likelihood of finding a trading counterparty is maximized when all orders rub shoulders. Thus — at least in theory — by “fragmenting” the order flow between different trading venues, ATSs pose a danger to both price discovery and liquidity.
The regulatory response to this danger is to build a framework of rules that, in essence, turns the various disparate markets into a “virtual” single market, so that no matter where a trader submits an order, it will interact with orders posted on all of the other markets.
Canadian regulators achieve this by requiring any marketplace that displays orders to anyone (which includes all but the confidential block matching systems) must make its order book publicly available through an information vendor. In addition, professional traders processing client orders must comply with “best execution” and “best price” obligations.
The duty of best execution is an adjunct to the dealer’s fiduciary duty to the client, and requires the dealer to ensure that each client receives that combination of price, speed, certainty and cost of execution that is most appropriate for each client. This is commonly thought to put an onus on the dealer (albeit of somewhat uncertain scope) to access pricing and depth of market information in alternative markets when processing client trades.
The best price obligation is more specific and is designed to prevent “trade throughs.” A trade through occurs when a lower priced bid, or higher priced ask is allowed to trade “through” (or ahead of) a superior bid or ask. This is a violation of the most basic principle of stock trading — that the highest priced bid (or lowest priced ask) have priority for execution over all other orders. At present, the “best price” obligation attaches to dealers, and requires them to put in place policies and procedures that minimize the likelihood that a trade through will occur. It is a duty owed to the entire marketplace and not merely to the dealer’s clients.
In practice, many dealers comply with this obligation by using a “smart order router” (SOR). A smart order router is connected to all trading venues, and will aggregate information on pricing and depth of market in all marketplaces. Because an order may be too large to be executed in the marketplace displaying the “national best bid or offer” (NBBO), the SOR uses complex algorithms to first direct part of the order to the marketplace displaying the NBBO, and the remainder to other markets — again on the basis of price priority. SOR services may be offered by a third party vendor, a marketplace or by the dealer itself using proprietary software.
The combination of order visibility, the duties of best execution and best price and the availability of SORs thus addresses the fragmentation problem by pooling price information and liquidity across the different marketplaces.
A downside of having multiple marketplaces, however, is that only price, rather than price-time priority can effectively be enforced given existing technology. Herein lies the problem. Exploiting the absence of inter-market price-time priority, some trading venues have created order types that pose a danger to the virtual single market.
Some marketplaces, for example, have allowed their customers to enter “pegged” orders that adjust automatically to match the NBBO. These marketplaces then allow these orders to be executed ahead of identically priced orders that were previously posted on another marketplace — a practice that is highly damaging to the marketplace as a whole. The current technological state-of-the-art does not allow regulators to generally enforce inter-market price-time priority. However, pegged orders result in an obvious violation of inter-market price-time priority — and one that they are in a very good position to enjoin and to enforce.
Allowing pegged orders to scoop the NBBO does more than create the impression of an unfair market. It allows traders using pegged orders to effectively remove their orders from the price discovery process. It also imprisons liquidity within a single marketplace, reducing the extent to which orders on different marketplaces interact. If my bid on Market A is the NBBO, for example, I would normally expect that a matching offer on Market B will be forwarded to Market A for execution. However, if Market B permits pegged orders, an inferior bid in Market B’s order book will jump the queue, leaving my order unexecuted. If this happens often, I will clearly think twice before lining Market A’s books — or any other market’s books — with orders.
This is simply because pegged orders reduce the returns to posting limit orders. This constitutes a direct assault on what makes stock exchanges tick. Those who post limit orders are liquidity makers, since they offer other traders the opportunity to trade at the posted price. Those who hit these orders are liquidity “takers.” Since liquidity is a valuable commodity, a limit order thus has an “option” value to all potential traders. It is for this reason that most modern stock trading venues actually pay traders to post limit orders, charging only the “active” side on any trade that results.
Liquidity makers and liquidity takers exist because traders and trading strategies are heterogeneous. One cannot exist without the other. Harming the interests of one harms the interests of both.
The use of parasitic orders such as pegged orders thus interferes with price discovery and the pooling of liquidity from different markets. It is also anti-competitive, because it penalizes traders who do not post their liquidity to the market (or markets) permitting pegged orders. Indeed, the very purpose of pegged orders may be to force liquidity out of some venues and into others. Since no market can survive without offering liquidity (i.e., having a book of limit orders), the widespread use of pegged orders is fundamentally inconsistent with a competitive marketplace.
In the extreme, permitting pegged orders potentially leads to a Doomsday scenario in which all markets imprison some or all of their liquidity by permitting pegged orders. This could potentially lead us back in the direction of a monopoly provider of trading services — to the detriment of all but the monopolist.
Some trading venues have started to use another strategy whose effects are similar to pegged orders. Many institutional and other large block traders use confidential trading venues (such as the upstairs market and large block ATSs) to keep their orders secret. This is designed to ensure that public knowledge of the order does not move the market price against the investor. By regulatory fiat, these orders must be executed at or between the best bid and ask.
Because these “crossing markets” or “dark pools” use a price derived from the public market, they contribute nothing to price discovery. Nonetheless, by giving big block traders better execution than they can achieve in the public market, they substantially enhance big block liquidity. It is for this reason that these dark pools are allowed to exist.
A practice has recently sprung up, however, of using dark pools to trade small orders that would normally be destined for the public market. Since these orders are executed at or between the best bid and ask, they have the potential to offer clients some measure of price improvement. However, like pegged orders they are “parasitic orders” that remove order flow from the visible public market, adversely impacting both price discovery and public market liquidity.
They are another device that can be used to imprison liquidity within a dealer and/or marketplace, to the ultimate long-term detriment not merely of the public market, but the dealer’s own clients. Unlike large block crosses, there is no rationale for withholding small trades from the public market, other than to allow the dealer to earn two commissions on the trade — and to withhold business from other dealers and/or marketplaces.
Canadian financial markets are at a critical juncture, with a variety of new trading venues aggressively challenging the hegemony of the historical incumbent. It is thus a particularly important time to make sure that the rules of the game ensure that competition, though vigorous, is fair. It is imperative that regulators take the bull by the horns and act quickly to prohibit trade types that operate against the best interest of Canadian corporations and the entire financial community.
